EIS Seminar Series

Abstract

Speaker Bio

In contrast to time and space complexity, the energy complexity of computation has received relatively little attention. However, recent work has shown that, for error control circuits implemented in VLSI, coding performance and energy are related in a non-trivial way. First, universal lower bounds on the energy complexity of all encoding/decoding circuits are derived. Polar codes are the first provably capacity achieving computationally efficient error control codes. However, in contrast to the universal lower bounds, lower bounds on the energy complexity of sufficiently high rate polar codes are derived, which are reachable up to polylogarithmic factors on a mesh network. These results illuminate a gap between achievable upper and lower bounds. As well, this work suggests fundamental tradeoffs in time, space, and energy for computation in general.

WHEN:   Monday, June 27, 2016 WHERE:   Porter Hall B34 TIME:   2:30 p.m.

Christopher Blake

Christopher Blake is a PhD student at the University of Toronto in the Department of Electrical and Computer Engineering. He received his undergraduate degree at the University of Toronto and his Masters at the Massachusetts Institute of Technology. His research interests include error control coding and the fundamental physical limits of computation and communication.

"On Scaling Rules for the Energy of Polar Encoders and Decoders"